Evaluation of simulated climatological diurnal temperature range in CMIP5 models from the perspective of planetary boundary layer turbulent mixing

Abstract

This study examines the effects of modeled planetary boundary layer (PBL) mixing on the simulated temperature diurnal cycle climatology over land in 20 CMIP5 models with AMIP simulations. When compared with observations, the magnitude of diurnal temperature range (DTR) is systematically underestimated over almost all land areas due to a widespread warm bias of daily minimum temperature (Tmin) and mostly a cold bias of daily maximum temperature (Tmax). Analyses of the CMIP5 multi-model ensemble means suggest that the biases of the simulated PBL mixing could very likely contribute to the temperature biases. For the regions with the cold bias in Tmax, the daytime PBL mixing is generally underestimated. The consequent more dry air entrainment from the free atmosphere could help maintain the surface humidity gradient, and thus produce more surface evaporation and potentially lower the Tmax. The opposite situation holds true for the regions with the warm bias of Tmax. This mechanism could be particularly applicable to the regions with moderate and wet climate conditions where surface evaporation depends more on the surface humidity gradient, but less on the available soil moisture. For the widespread warm bias of Tmin, the widely-recognized overestimated PBL mixing at nighttime should play a dominant role by transferring more heat from the atmosphere to the near-surface to warm the Tmin. Further analyses using the high resolution CFMIP2 output also support the CMIP5 results about the connections of the biases between the simulated turbulent mixing and the temperature diurnal cycle. The large inter-model variations of the simulated temperature diurnal cycle primarily appear over the arid and semi-arid regions and boreal arctic regions where the model differences in the PBL turbulence mixing could make equally significant contributions to the inter-model variations of DTR, Tmax and Tmin compared to the model differences in surface radiative processes. These results highlight the importance and need for accurate descriptions of the PBL processes with respect to the turbulent mixing in order to improve the temperature diurnal cycle simulations in climate models.